AbstractMicrogrids are a promising means to broadly integrate distributed energy resource (DER) systems. Seen as an enabling technology within smart grid infrastructure, microgrids offer solutions for both customers and system operators of future power networks. However, the optimal management and control of many generation, load, and storage assets local to a microgrid presents as a daunting challenge. In this paper, an agent-based control architecture for microgrids, capable of coordinating and cooperatively achieving multiple user-defined objectives in real-time is presented. Herein, the authors present fundamental properties required by a decentralized microgrid multi-agent system (MAS), define agent capabilities that support resilient operations, and discuss aspects of the microgrid MAS necessary to smoothly transition between system states. The microgrid MAS developed allows control decision-making to be shifted to a lower hierarchical level, capitalizing on the ability of distributed agents to self-order and cooperate towards an optimal power management solution, given assigned goals. As smart grid technologies continue to evolve, the microgrid operational concepts discussed, along with robust power management and control issues, will play an increasingly significant role.